We develop and apply ‘wet’ molecular biology and ‘dry’ computational approaches with the goal of extracting the most information possible from the smallest sample input at the lowest cost and then turn that information into actionable knowledge. To achieve this, we implement multiplexed sequencing approaches to assay circulating DNA and RNA from biofluid samples (i.e., ‘liquid biopsies’), as well as statistical methods (including network approaches) for integrating omics data, functional annotations, and patient information to improve disease state prediction, particularly for small (including n=1) and heterogeneous cohorts. These approaches also aid in identifying the optimal preclinical model, and interpretation of resulting data, to better capture patient-relevant signatures.